#4: Tensor in TensorFlow
The video shows how to create a tensor and perform basic operations in TensorFlow.
Timeline
(Python 3.7.12; TensorFlow 2.8)
00:00 - Begin
00:09 - Outline of video
00:22 - Open notebook in Google Colaboratory
01:28 - Create a Rank 0 tensor: tf.constant()
02:42 - Create a Numpy array
03:20 - Create a Rank 1 tensor
04:40 - Create a Rank 2 tensor
05:39 - Create a Rank 3 tensor
07:57 - Convert a tensor to Numpy array
09:40 - Basic math operations
10:00 - Basic math operations: Create two tensors: x, y
10:38 - Basic math operations: add: tf.add(x,y)
11:05 - Basic math operations: subtract: tf.subtract(x,y)
11:53 - Basic math operations: multiply: tf.multiply(x,y)
12:47 - Basic math operations: matmul: tf.matmul(x,y)
14:01 - Basic math operations: divide: divide(x,y)
14:35 - Basic math operations: add: x + y
14:42 - Basic math operations: subtract: x - y
14:55 - Basic math operations: divide: x / y
15:03 - Basic math operations: floor division: x // y
15:16 - Basic math operations: matmul: x @ y
15:19 - * * * Correction: Misspoke 'ampersand' for '@' instead of 'at sign'
15:55 - Operations: create a tensor
16:56 - Operations: max value: tf.reduce_max(x)
18:20 - Operations: max value by axis: tf.reduce_max(x, axis=0)
20:30 - Operations: get index for largest value: tf.argmax(x)
23:11 - Operations: softmax: tf.nn.softmax(x)
25:52 - Shape and rank: create a rank 4 tensor: tf.zeros()
29:32 - Shape and rank: x.dtype
29:57 - Shape and rank: x.ndim
30:21 - Shape and rank: x.shape
30:42 - Shape and rank: number of elements along axis=0: x.shape[0]
31:33 - Shape and rank: number of elements along last axis, axis=-1: x.shape[-1]
31:52 - Shape and rank: total number elements: tf.size(x)
32:33 - Ending notes